Final Activity Report Summary - NEUROGRASP (Decoding of grasping movements from frontal and parietal cortex for the development of a neural prosthesis)
The way we use our hand plays a central role for our interactions with the environment and for what defines us of being human. Many brain areas are involved in the generation of hand movements, but two regions in the premotor and parietal cortex play a particular important role for the initial, or high-level, planning of hand grasping movements. The generation of these initial movement intentions involves the transformation of visual information into a motor plan that can be further processed and executed.
The execution of such movements is disrupted in paralysed patients, e.g. patients with severe spinal cord injury or stroke, but the planning capacity of the brain for these movements is still preserved. While neural commands can no longer reach the hand muscles, it may be possible to artificially read out and interpret these brain signals and deliver them to either the natural or a prosthetic hand - if one would understand how hand planning signals are generated and represented in the brain.
In this project, we investigated in macaque (rhesus) monkeys how intentions to move the hand are generated and represented in the brain, and how these signals can be read out in real-time. Animals were trained to perform a delayed hand grasping task, in which an object can be grasped with different grasping behaviours (e.g. precision grip or power grip) and in various hand orientations and grasp target positions. Animals were rewarded for each successful execution of the task with small amount of juice. Once the animal has learned the task, neural activity was recorded from two brain areas, the anterior intraparietal area (AIP) and the area F5, which both generate early hand grasping instructions. By correlating the neural responses with each other and with the animal's behaviour, we found that high-order grasping plans are encoded in AIP and F5 and that these areas interact with each other specifically during the preparation of these movements. In a second set of experiment, brain signals from these areas were decoded in real-time by a computer, while an animal planned, but not executed, hand grasping movements. This provided proof of concept that hand grasping movements can be predicted from these high-order brain areas in real-time.
For these experiments, the use of macaque monkeys is essential for several reasons. First, only primates share with humans the dominant role of the visual system, the hand dexterity, and the way objects are grasped. Other species do not have this hand dexterity. Second, since the ultimate goal of understanding the hand movement system is to develop a neural prosthesis for human patients, studying the neural control of hand grasping in non-human primates is a necessary and appropriate first step.
The execution of such movements is disrupted in paralysed patients, e.g. patients with severe spinal cord injury or stroke, but the planning capacity of the brain for these movements is still preserved. While neural commands can no longer reach the hand muscles, it may be possible to artificially read out and interpret these brain signals and deliver them to either the natural or a prosthetic hand - if one would understand how hand planning signals are generated and represented in the brain.
In this project, we investigated in macaque (rhesus) monkeys how intentions to move the hand are generated and represented in the brain, and how these signals can be read out in real-time. Animals were trained to perform a delayed hand grasping task, in which an object can be grasped with different grasping behaviours (e.g. precision grip or power grip) and in various hand orientations and grasp target positions. Animals were rewarded for each successful execution of the task with small amount of juice. Once the animal has learned the task, neural activity was recorded from two brain areas, the anterior intraparietal area (AIP) and the area F5, which both generate early hand grasping instructions. By correlating the neural responses with each other and with the animal's behaviour, we found that high-order grasping plans are encoded in AIP and F5 and that these areas interact with each other specifically during the preparation of these movements. In a second set of experiment, brain signals from these areas were decoded in real-time by a computer, while an animal planned, but not executed, hand grasping movements. This provided proof of concept that hand grasping movements can be predicted from these high-order brain areas in real-time.
For these experiments, the use of macaque monkeys is essential for several reasons. First, only primates share with humans the dominant role of the visual system, the hand dexterity, and the way objects are grasped. Other species do not have this hand dexterity. Second, since the ultimate goal of understanding the hand movement system is to develop a neural prosthesis for human patients, studying the neural control of hand grasping in non-human primates is a necessary and appropriate first step.